Salt Pollution of the Middle and Lower Sections of the River Werra (Germany) and Its Impact on Benthic Macroinvertebrates

Limnologica 41 (2011) 113–124

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

Contents lists available at ScienceDirect provided by Elsevier - Publisher Connector

Limnologica

journal homepage: www.elsevier.de/limno

Salt pollution of the middle and lower sections of the river Werra (Germany) and its impact on benthic macroinvertebrates

Ulrich Braukmann a,∗, Dirk Böhme b

a Universität Kassel, Fachgebiet Gewässerökologie/Gewässerentwicklung, Nordbahnhofstraße 1a, 37231 Witzenhausen, Germany b BWWU Büro für Wasserwirtschaft und Umwelt, Blümnerstraße 18, D-04229 Leipzig, Germany

article info abstract

Article history: In two survey phases (2003 and 2008) organic, nutrient and salt contamination parameters have been Received 22 June 2010 investigated in the lower Werra in order to estimate the importance of these different kinds of pollution Accepted 16 September 2010 for the quality component of macroinvertebrates according to the European Water Framework Directive. The chemical and biological investigations have been carried out comparing a “reference” section with- Keywords: out salt contamination with the salt contaminated section due to the potash mining industry from Vacha Werra organic to Hannoversch Münden close to the mouth of the Werra. The results show that the drastic differences Nutrient between the macroinvertebrate assemblages of the Werra upstream and downstream the salt contami- Salt contamination Potash mining nated sections are clearly caused by the salt load. The other kinds of chemical impacts are not responsible Chemical for the observed fundamental change within the composition of the benthic invertebrate assemblage. Biological assessment General degradation of stream morphology, indicated by macroinvertebrates, shows a good ecological Macroinvertebrates status for the non-salt-contaminated part of the river and a bad status for the salt contaminated sites of European Water Framework Directive the lower Werra. © 2010 Elsevier GmbH. All rights reserved.

Introduction and 50◦ Gh for the total water hardness. For the third time the limit was adjusted for total hardness to 65◦ Gh in 1996. A fourth time in The River Werra has been polluted by waste waters with 1998/1999 the limits were fixed again to a higher level of 90◦ Gh extremely high salt concentrations from the potash mining indus- which is permitted up to now. try between the Thuringean sampling sites Vacha and Gerstungen Both the extremely high salt load as permitted and the specific for more than 100 years. This salt contamination has continued up salt composition, differing significantly from natural ionic compo- to now downstream not only to the Hessian section of the Lower sition of riverine or estuary waters, form a specific situation among Werra but also to the river Weser until its mouth. the rivers not only in Germany but also in central Europe. The saline At the beginning of the 20th century the potash mining industry pollution leads to a permanent chemical impact on the aquatic emitted salt waste water for the first time in 1901 into the Werra coenoses. (Hübner 2007). Just 10 years later salt pollution caused the first The European Water Framework Directive (EU WFD 2000) clas- problems for the drinking water supply of the city of Bremen. In sifies the ecological status of running waters into 5 quality classes: 1913 a potash commission stated first threshold values for impor- 1 = high, 2 = good, 3 = moderate, 4 = poor, 5 = bad (EU WFD 2000). tant salt water components. For the Werra at the sampling site It demands that natural bodies of water attain a good ecolog- Gerstungen the thresholds for chloride amounted to 842.5 mg/l, ical status by 2015 and heavily modified water bodies a good for total hardness to 48.4◦ German hardness (Gh), for the Weser ecological potential. The ecological status or potential is defined at Bremen to 250 mg/l for chloride and to 20◦ Gh for total hard- by multimetric index systems for the biotic quality components: ness. A first adjustment of these limits to rising amounts of salt fish, macrozoobenthos, macrophytes and phytobenthos (EU WFD effluents took place in 1924 with values of 1781 mg/l chloride and 2000). 63.1◦ Gh at Gerstungen. For the river Weser at Bremen the limits The main objective of this study is to analyse the specific were raised to 350 mg/l for chloride and 23◦ Gh for total hardness. pollution of the Werra and to work out the importance of salt con- A second adjustment fixed a threshold of 2500 mg/l for chloride tamination compared to other anthropogenic impacts. This will be done using the example of the benthic macroinvertebrate communities in the salt polluted section of the Werra, compared to non salt ∗ polluted reference sections of the River Werra and to a naturally Corresponding author. Tel.: +49 5542 98 1632. E-mail address: [email protected] (U. Braukmann). salty stream, the Salzbach near Witzenhausen.

Fig. 1. Sampling sites at the River Werra.

In a first step, the macrozoobenthos will be evaluated using eco- sections of the Werra have been classified as type 9.2, a large logical metrics according to the German WFD methodology and highland river with a large floodplain (over 300 m wide). All river the results will be related to the saline pollution. Further the ben- sections under study belong to this type. The Salzbach is a small thos data as well as environmental parameters will be analysed tributary of the River Werra near Witzenhausen which may be applying a multivariate approach in order to estimate the relative classified morphologically as a type 7, a small coarse substrate dom- importance of saline pollution, physical habitat, and nutrients for inated calcareous highland stream according to Pottgiesser and the community structure. Sommerhäuser (2004). Its high natural salt concentration results from geological layers of zechstein in its catchment area. Because Study sites of its extremely high salinity the Salzbach stream chemically and biologically represents a special type which is not comparable to The study sites cover the natural geographic regions of the mid- other “regular” streams of type 7. dle and lower section of the River Werra. Based on a German system Fig. 1 shows the position of the actual potash mines and the of stream types (Pottgiesser and Sommerhäuser 2004, 2008) both sampling sites of the sampling periods in 2003 and 2008. U. Braukmann, D. Böhme / Limnologica 41 (2011) 113–124 115

Table 1 Sampling site parameters: site code, mean discharge, distance from source, mean slope of river bed, exposition to saline pollution, sampling frequency in 2003 and 2008. W: River Werra; S: creek Salzbach.

Site code Sampling site Distance from Slope of river bed Mean discharge Exposition to salt Sampling Sampling source [km] [‰] [m3/s] pollution frequency 2003 frequency 2008

Chemistry Biology Chemistry Biology

W01 Breitungen 101 0.85 21.1 No 2 1 W02 Barchfeld 106 0.80 21.2 No 4 3 2 1 W03 Bad Salzungen 113 0.20 21.5 No 1 W04 Vacha 129 0.35 22.1 Yes 1 W05 Harnrode 137 1.49 28.3 Yes 1 W06 Widdershausen 145 0.35 29.1 Yes 1 W07 Gerstungen 158 0.45 31.0 Yes 4 3 2 1 W08 Hörschel 174 0.36 35.0 Yes 1 W09 Frieda 232 0.69 41.4 Yes 6 3 1 W10 Niederhone 243 0.65 43.2 Yes 6 3 1 W11 Albungen 245 0.70 47.4 Yes 6 3 1 W12 Witzenhausen 275 0.81 49.0 Yes 6 3 1 W13 Blickershausen 282 0.37 50.4 Yes 6 3 2 1 W14 Hann. Münden 297 0.34 51.0 Yes 1 S01 Wendershausen 0,2 79.89 0.01 No 6 3

Methods in salt, the Salzbach, was investigated in order to compare a natural salt water stream with an anthropogenically salinised stream and Sampling design the structure of their benthic invertebrate coenoses.

Survey phase 1, 2003 Survey phase 2, 2008 In order to evaluate changes in the zoocoenosis of macroinver- In February and August 2008 a second phase of investigations tebrates by the impact of salt effluents, the Werra was investigated was established in order to rate the actual impact of different pol- in 2003 upstream from the first-known salt contamination area at lutants of the lower Werra downstream. In this phase the impact Barchfeld (Middle Werra). This sampling site is located upstream of different pollutants (organic oxygen consuming waste water from the waste water treatment plant (WWTP) at Bad Salzungen from domestic sources, plant nutrients and salt components) on to avoid influences of domestic sewage. the Werra downstream was studied. In February 11 sites were The sampling site near Gerstungen (Middle Werra) is located investigated by water chemistry analyses regarding the prevail- in the section of the highest salt contamination in this river. The ing point sources of salt contamination as well as the areas of salt immissions originate not only from waste pipes but also from diffuse intrusions of salt into the Werra due to salt disposal into diffuse intrusions of salt waste water from the underground, caused the underground into the “Plattendolomit”. Two sites were estab- by the rising salt enriched groundwater. This salty waste water has lished in the non-salt-polluted segment of the Werra between been anthropogenically pressed into the geological formation of Breitungen and Barchfeld. The 9 other sites reveal the increasing the so called “Plattendolomit”, a special form of zechstein, formerly salt-contamination and the following dilution by the tributaries of believed to be a dense salt-retaining layer several hundred meters the Werra. In August chemical and biological studies were carried below the surface. Due to numerous tectonic fracture zones this out at 4 sites, two in the non-salt-polluted (Breitungen and Barch- layer became locally permeable for rising salt waste waters. feld) and two in the salt contaminated section (Gerstungen and The following 4 sampling sites are located downstream in Bickershausen, see Fig. 1). Table 1 is a compilation of some general the section of the Lower Werra at Frieda, Niederhone, Albungen, data characterising the investigated sampling sites and the sam- Witzenhausen and Blickershausen. pling design of this study which were used for statistical analyses. In 2003 chemical surveys took place 4 times in the Werra at Chemical analyses covered water temperature, oxy- Barchfeld and Gerstungen, 6 times from Frieda to Blickershausen gen content and saturation, electrical conductivity, pH, and in the Salzbach at Wendershausen. Chemical sampling data orthophosphate–phosphorus, nitrate–, and ammonium–nitrogen, cover all seasonal aspects from spring to winter. At the same sites total nitrogen, hydrogen carbonate, total hardness, carbonate morphological and biological surveys were carried out three times hardness, chloride, sulphate, sodium, potassium, calcium and in spring, summer and autumn to estimate the ecological effects on magnesium. In 2008 biochemical oxygen demand in 5 days (BOD5) benthic invertebrates of the Lower and Middle Werra. All sections of and dissolved organic carbon (DOC) were analysed additionally. the river Werra were investigated upstream from the confluence of Temperature, oxygen, pH and electrical conductivity were mea- tributaries in order to avoid their diluting influences on the Werra sured in the field using measuring equipment by WTW “MultiLine invertebrate coenoses. Additionally, a small stream naturally rich P4”.

Table 2 Mean percentage of substrate types at sampling sites in 2003 (according to Hübner 2007). phytal: plants; pelal: mud, silt; psammal: sand; akal: gravel; lithal: larger stones.

Site Litho-phytal Xylal Macropelal Micropelal Psammal Akal Mesolithal Macrolithal

Barchfeld 5 5 10 2 28 8 22 20 Gerstungen 3 2 17 18 37 17 Frieda 2 13 63 18 3 Niederhone 2 12 13 20 50 Albungen 3 13 2 3 28 43 Witzenhausen 5251552 Blickershausen 10 5 43 42 Salzbach Wendersh 3 8 22 28 33 5 116 U. Braukmann, D. Böhme / Limnologica 41 (2011) 113–124

Fig. 2. Chemical data of the River Werra and the creek Salzbach in 2003. U. Braukmann, D. Böhme / Limnologica 41 (2011) 113–124 117

Fig. 3. Biological data of the Werra and the creek Salzbach in 2003: (a) number of benthic invertebrate taxa, (b) macroinvertebrate diversity, (c) dominance values of 6 dominant taxa in the salinised Werra, (d) biological similarity of all sites, cluster analysis of site-pooled macrozoobenthos data. W: Werra; W02: Barchfeld; W07: Gerstungen; W09: Frieda; W010: Niederhone; W11: Albungen; W12: Witzenhausen; W13: Blickershausen; S01: Salzbach at Wendershausen.

All other chemical parameters were analysed according to Ger- mainly interpreted as an indicator of physical habitat deficit. How- man Standard Methods (DIN). ever, it is sensitive to most other anthropogenic impacts too (Lorenz et al. 2004). This fact is crucial for identification of key stressors, Invertebrate sampling followed by decisions on priority strategies in river restoration. Fur- ther analyses in this paper will focus on the benthos community For biological sampling of macroinvertebrates a handnet structure. according to Braukmann (2000) was used, details see Hübner Hierarchical Cluster analysis (UPGMA, Sneath and Sokal 1973) (2007). The collection was performed as multihabitat sampling of was used in Q-mode to visualise the faunistic similarity structure of 20 m length. It covered 1 m2 in total and consisted of 10 subsamples the macrozoobenthos communities. For this analysis the seasonal of 0.1 m2 each. All substrate types were collected corresponding to samples were pooled by sampling sites to give an overall view of their relative abundance in the section in question. the community composition. Further the impact of seasonality on Organisms found were analysed alive and counted in the field. the community structure was assessed by ANOSIM (Clarke 1993) Unidentifiable taxa were sorted from the sample, preserved in 70% using unpooled data. Finally, partial Mantel tests (Mantel 1967; ethanol and identified in the laboratory, if possible to species level. Legendre and Legendre 1998) on multiple environmental variable Abundance was expressed in individual numbers per m2. groups were carried out to rank the importance of longitudinal zonation, saline pollution, nutrient pollution/eutrophication and Statistical data analysis physical habitat structure for the benthic zoocoenoses. The selec- tion and assignment of abiotic variables into these four groups is Descriptive statistics documented in the lower part of Table 6. As the raw benthos data matrices contain few dominant taxa Box and whisker plots of original abiotic variables but many zeros and singleton records, the Steinhaus index was According to the German evaluation system “Perlodes” (Hering calculated from raw abundance data after log(x + 1) transforma- et al. 2004), 4 metrics were calculated from site-pooled macro- tion as recommended by Legendre and Legendre (1998). This index zoobenthos data (German Fauna Index (Lorenz et al. 2004), % is asymmetric (double absences are not counted as indicators of metarhithral taxa (Moog 1995; Schmedtje and Colling 1996), % EPT proximity) and gives more weight to common presence of species. (Richardson 1928), Diversity (Shannon-Wiener, Margalef, Simp- Similarity matrices based in the Steinhaus index were used in all the son)). above described analyses. The environmental data matrices (inde- These metrics are aggregated to an index of “Overall degrada- pendent variables and covariables) were normalised, followed by tion”. This index is a compulsory part of the “Perlodes” system and is calculation of Euclidean distances. 118 U. Braukmann, D. Böhme / Limnologica 41 (2011) 113–124

Table 3 Biological assessment of the River Werra in 2003 by macroinvertebrates with the programme “Asterics/Perlodes (2008)” (according to Hering et al. 2004). E: Ephemeroptera, P: Plecoptera, T: Trichoptera; quality classes: 1: high, 2: good, 3: moderate, 5: bad; –: no adequate stream type existing (unpolluted zechstein creek).

Sampling sites German Fauna Index % Metarhithral taxa EPT Diversity General degradation

Barchfeld 2 3 3 1 2 Gerstungen 5 5 5 5 5 Frieda 5 5 5 5 5 Niederhone 5 5 5 5 5 Albungen 5 5 5 4 5 Witzenhausen 5 5 5 5 5 Blickershausen 5 5 5 5 5 Salzbach Wendershausen – – – – –

The Mantel analysis is based on the assumption that correlation underground (dolomite-layer), or as solid waste on salt piles. The between (dis)similarity structure of samples in biotic space and of salt from the piles are washed out into the Werra by precipitation. sampling sites in environmental space may be used as an indicator In contrast to nutrients except from carbonate hardness all salin- for causal relationship between the biota and the environmental isation parameters (electrical conductivity, carbonate hardness, parameters under study. If a strong statistical relation is detected, chloride, sodium, calcium, magnesium and sulphate) showed clear causality may be discussed using additional ecological and phys- differences between the two uppermost sites in the non salinised iological knowledge. The partial Mantel test may be seen as the sections at Barchfeld and the salinised site at Gerstungen and all multivariate equivalent to the partial correlation coefﬁcient. The other downstream salinised Werra sites. degree of relationship between two distance matrices is calculated In sharp contrast to the Werra sites especially magnesium and after the effect of a third matrix has been removed, i.e. the effect of potassium values are much lower in the naturally salty Salzbach. At one single subsets of variables can be separated by partialling out this site with the highest electrical conductivity and sodium con- the effect of the others. centration and with the second highest chloride concentration of all Statistical analyses were conducted using Statistica, version 6.0 sites, very low magnesium (mean 31.2 mg/l) and potassium (mean (descriptive statistics), MVSP 3.1 (clustering) and Brodgar 2.5.1 11.9 mg/l) concentrations were measured which were much more (ANOSIM, partial Mantel tests). comparable to the non salinised Werra section (Mg mean 13.8 mg/l, K mean 4.2 mg/l at Barchfeld) but to the salinised sections of the Results Werra (Fig. 2f and g). The median magnesium concentration at Barchfeld in 2003 was Survey phase 2003 15 times lower, the median potassium concentration even 35 times lower than at the site at Gerstungen. In 2003 besides morphological parameters like substrate dis- tribution of the river bed chemical parameters showing nutrient Benthic invertebrates and salinisation characteristics of the Werra and the Salzbach were Benthic macroinvertebrate assemblies changed drastically studied. between the non salinised Werra section and the salinised one. This is illustrated by the number of invertebrate taxa (Fig. 3a), different Substrates diversity indices (Fig. 3b) and the most dominant taxa (Fig. 3b). All sites of the Werra were sampled close to river edge. They Some results of the assessment of benthic invertebrate showed rather similar substrate conditions with prevailing coarse assemblies carried out using the programme “Asterics/Perlodes” substrates (macro-, mesolithal, akal) which are compiled in Table 2. developed for bioindication by (benthic invertebrates in Germany according to WFD (Hering et al. 2004) are compiled in Table 3. Chemistry This table shows typical aspects of the metric “general degrada- Fig. 2 is a compilation of several chemical parameters which tion” which is designed as a biological feature indicating mainly were regarded important for ecological characterisation and biocenotically relevant morphological degradations. assessment of the River Werra and, for comparison, a small nat- The metric “general degradation” is derived from parameters urally salty stream, the Salzbach. like “German Fauna Index”, “% Metarhithral taxa”, “EPT” (proportion of ephemeroptera, plecoptera and trichoptera as important Nutrients parts of the total invertebrate fauna) and diversity aspects. The investigated nutrient parameters total nitrogen and ortho In the over-all aspect “general degradation” the sampling site phosphate phosphorous showed similar ranges at all Werra sites at Gerstungen appears in a good quality status (class 2), the other (Fig. 2a and b). Median values of total N lay between 3.8 an 4.8 mg/l. two sites, clearly due to the salt impact, are classiﬁed bad (class In the Thuringean section at Barchfeld and Gerstungen. Phospho- 5). In detail also the metrices “German Fauna Index” and “diver- rous values were higher in range and median (median between 0.17 sity” show class 2 (good) for the non-salt-contaminated sites, and 5 and 0.19 mg/l P) than in the downstream Hessian sections (median (bad) for the contaminated ones. The metrices “percentage of meta- between 0.1 and 0.12 mg/l P). rhithral taxa” and “proportion of EPT” (ephemeroptera, plecoptera, trichoptera) indicate also class 1 (high status) for Breitungen, class Salinisation parameters 3 (good) for Gerstungen and class 5 for the salt polluted sections For the characterisation of salt load the most important com- Gerstungen and also Blickershausen near the mouth of the Werra. ponents of the marine salt deposits of the geological zechstein formations like sodium, potassium, calcium, magnesium, chlo- Survey phase 2008 ride and sulphate were analysed. Magnesium and potassium are extracted by the potassium industries and processed further as Chemistry components of agricultural fertiliser. The other salts are dumped as Additional to nutrient and salinisation parameters in 2008 BOD waste water, either directly into the Werra, or recharged into the and DOC were measured in order to obtain an impression of the U. Braukmann, D. Böhme / Limnologica 41 (2011) 113–124 119

Fig. 4. Chemical data of the river Werra in 2008, (a–f): February, (g and h): August. actual organic impacts from domestic waste water regarding the for the transition from good to moderate quality for large mountain question if the fact that less than 50% of Thuringian inhabitants rivers (type 9.2) (Fig. 4a). Despite the lower degree of connection to in the Werra catchment area are connected to WWTP leads to an domestic waste water treatment plants of townships in Thuringia impact on benthic invertebrates comparable to salinisation param- (50%), compared to Hessia (above 90%), BOD5 as well as DOC values eters. In Fig. 4 essential chemical parameters are compiled. were lower in Thuringian river sections than in Hessian ones. Ammonium concentrations as an indicator for insufficiently Organic pollution treated waste water were well below the orientation threshold + value of 0.3 mg/l NH4 –N for the transition from good to moderate The degree of organic pollution is indicated by BOD5 as well as by DOC (Fig. 4a). Also ammonium–nitrogen (Fig. 4b) indicates organic quality according to LAWA (2007) (Fig. 4b). pollution. In order to characterise organic pollution, four sites (Bre- itungen, Barchfeld, Gerstung and Blickershausen) were sampled in Nutrients August 2008 during the period of pessimal water quality condi- As in 2003 nitrate–N and orthophosphate–P were also measured tions regarding the oxygen conditions in rivers in order to find out in 2008. In February, during a slight flood period, nitrate–N values the maximal impact of organic pollution on benthic invertebrates. were similar at all stations (Fig. 4c). The maximal value reached Except Blickershausen at all other sites BOD values clearly lay below 4.2 mg/l at Blickershausen. In August Nitrate–N values lay between the orientation value of LAWA (2007) of 6 mg/l as a threshold value 3.0 and 3.3 mg/l. 120 U. Braukmann, D. Böhme / Limnologica 41 (2011) 113–124

Fig. 5. Biological data of the river Werra in 2008: (a) number of benthic invertebrate taxa, (b) macroinvertebrate diversity, (c) dominance values of 6 most dominant taxa in the salinised Werra, (d) Saprobic index of 6 sites.

Downstream from Gerstungen orthophosphate values from formations. Here the chloride concentration was 25 mg/l. At Ger- February samples lay above the orientation value of 0.07 mg/l stungen 2037 mg/l were measured, a value which was 81 times according to LAWA (2007) (Fig. 4d). The values in Thuringia were higher than at Barchfeld. The official “legal” threshold of 2.500 mg/l lower than in Hessen despite the low degree of waste water chloride in the Werra which exists since 1942 is about 12.5 treatment. In August 2008 phosphate–P lay between 0.15 mg/l times higher than the orientation value of 200 mg/l between a (Blickershausen, Hessen) and 0.26 mg/l (Barchfeld, Thuringia). As good a moderate ecological status for running waters according in the case of organic load also in the nutrient load no evident to LAWA (2007). Similarly drastic differences were also found in differences between the sampling sites could be observed either. total hardness (mainly due to wasted magnesium), magnesium and potassium concentrations (Fig. 4f and h). Salinisation parameters Carbonate hardness (mainly determined by calcium bicarbon- In 2008 the same salinisation parameters were analysed as in ate) stays nearly stable along the river, total hardness (dominated 2003. by magnesium discharged into the river) increases abruptly downstream from the first salt influent at Vacha reaching a maximum at Gerstungen. Electrical conductivity It is obvious that in contrast to the organic and nutrient load In the middle Werra between Barchfeld and Gerstungen the change in chemism of the Werra is clearly caused by all named salt electrical conductivity (Fig. 4c) increases drastically due to the components. The natural relation between concentrations of sev- direct discharge of salt waste water from the mining industry in eral ions especially between calcium, magnesium and potassium the area of Hattorf. Additionally the diffuse impact of upwelling typical for natural rivers, e.g. at Breitungen and Barchfeld above saltwater being pressed out of the dolomite layer by the waste salt the first salt influence, is dramatically inverted in the salty section brine which had been sunk into this geological layer affects the of the river beginning at Vacha. water quality of the river. Contrary to the organic and nutrient parameters the investi- gation of the salt components in 2008 along the sampling sites displayed a continuous increase of the electrical conductivity from Benthic invertebrates Breitungen with about 460 ␮S/cm rising at Vacha with the first salt Fig. 5 gives an overview of some aspects of benthic macroin- impacts up to the maximum of 7240 ␮S/cm at Gerstungen. Tribu- vertebrates found in August 2008 in non salinised and salinised taries dilute the electrical conductivity down to 2400 ␮S/cm at the sections of the river Werra. end of the Werra in Hannoversch Münden. Like in 2003 a considerable breakdown in taxa number and diversity was observed from the non salinised sections at Bre- Chloride, total hardness, magnesium, potassium itungen and Barchfeld downstream to the salinised section at At Barchfeld the Werra is a typical carbonate river with nat- Gerstungen. At Gerstungen just two taxa (Gammarus tigrinus and urally rather high values of magnesium and total hardness caused Potamopyrgus antipodarum reached 99.8% of the total number of all by geological conditions of the catchment area containing dolomite specimen over all taxa (Fig. 5c, Table 8 in Appendix). U. Braukmann, D. Böhme / Limnologica 41 (2011) 113–124 121

Table 4 Biological assessment of the River Werra in 2008 by macroinvertebrates with the programme “Asterics/Perlodes (2008)” (according to Hering et al. 2004). E: Ephemeroptera, P: Plecoptera, T: Trichoptera; quality classes: 1: high, 2: good, 3: moderate, 5: bad.

Sampling sites German Fauna Index % Metarhithral taxa EPT Diversity General degradation

Breitungen 2 1 1 2 2 Barchfeld 2 3 3 2 2 Gerstungen 5 5 5 5 5 Blickershausen 5 5 5 5 5

Table 5 Results of ANOSIM analysis.

ANOSIM 1-way permutation test: 9999 of 66512137 total permutations, global statistic = 0.179, p = 0.012*, nMDS stress = 0.0099

Group 1 Group 2 Max. permutations Statistic p

Spring Summer 1716 0.310 0.005** Spring Autumn 1716 0.235 0.017* Summer Autumn 1716 0.071 0.174

Levels of signiﬁcance: *0.05 ≥ p > 0.01, **0.01 ≥ p > 0.001, *** p ≤ 0.001.

Regarding organic pollution by means of saprobic indicators of tational procedures in the Mantel tests. The nMDS coupled to the macroinvertebrates, Breitungen reveals the lowest degree of pol- ANOSIM procedure gives a good two-dimensional representation of lution with a good quality status. The other sites are in a moderate the original similarity relationships within the resemblance matrix state. The differences between all sites are rather small which cor- (stress = 0.0099). responds well to the chemical data of organic and nutrient loads. Design and results of partial Mantel tests are presented in The assessment of biologically indicated general (mainly Table 6. As the river section between W02 and W13 covers a dis- morphologically induced) degradation by “Asterics/Perlodes” tance of about 176 km, first the importance of longitudinal zonation according to WFD (Table 4) results for 2008 are very similar to those was tested. Here only a slight relevance was found. However, lon- for the year 2003 (see Table 3). gitudinal variables were used as covariables within all further tests Again benthic invertebrates show a good ecological quality to fix the results for effects due to longitudinal zonation of the river. for the metric “general degradation” for the non salinised Werra The four variable groups used as explanatory variables section and, in all aspects, a bad quality for all sites within the in the partial Mantel tests No. 1–4 are relevant for the salinised Werra section. For the naturally salty creek Salzbach no community structure in the following order: salinity > habitat natural reference type exists, because this stream type is very rare structure > longitudinal zonation > nutrients and eutrophication. in Germany. Therefore using “Asterics/Perlodes” no result could be The last-mentioned impact seems to have nearly no relevance obtained. In this small stream neither morphological nor chemical under the paramount anthropogenic salinisation. pollutants could be observed. So this stream might be regarded as a reference stream for this very special stream type and give an idea of natural concentrations of magnesium or potassium at a Discussion high level of chloride. In this chapter the ecological effects of the chemical load of the Statistical analysis river Werra caused by the potash mining industry mentioned above Cluster analysis of site-pooled macrozoobenthos data from 2003 on the quality component macrozoobenthos of the EU-WFD will be (Fig. 3d) reveal a group formed by the salt polluted sites W07 to discussed. Caused by the far reaching changes of the hydrochem- W13. This group is well separated from the unpolluted site W01 and ical conditions as a consequence from the impact of salt wastes, from the natural saline site S01, although the latter has a magnitude a distinct change in the aquatic zoocoenosis takes place between of chloride concentration and conductivity similar to the polluted the non-salt-influenced upper section of the Werra and the salt Werra sites. Within the group of polluted sites, W09, W10, W12 and contaminated lower Werra below Hattorf. W13 form a compact sub-cluster indicating a vast loss of species Reasons for this change may generally be: salt, plant nutrient diversity in the most polluted river section. from agriculture and domestic waste water, organic pollution and Macrozoobenthos data were sampled in 2003 during three sea- hydromorphological deficits. sons (spring, summer and autumn). ANOSIM of non-pooled data From the chemical results presented in the former chapter it indicate that at least the spring samples differ considerably from could be concluded that salt pollution is the most important factor the other two seasons (Table 5). Thus it seems reasonable to use responsible for the biological changes. This hypothesis is clearly the sampling season as stratifying nominal variable within permu- supported by the outcome of the partial Mantel tests.

Table 6 Design and results of partial Mantel tests.

Partial Mantel test No. Matrix of explanatory variables Matrix of covariables Test statistic, rM p 1 Longitudinal zonation Habitat structure, nutrients and eutrophication, salinity 0.0770 0.1497 2 Habitat structure Longitudinal zonation, nutrients and eutrophication, salinity 0.4436 0.0050** 3 Nutrients and eutrophication Longitudinal zonation, habitat structure, salinity −0.0379 0.7264 4 Salinity Longitudinal zonation, habitat structure, nutrients and eutrophication 0.7759 0.0001***

Levels of signiﬁcance: *0.05 ≥ p > 0.01, **0.01 ≥ p > 0.001, *** p ≤ 0.001. Season was used as blocking variable for permutation in all tests. Longitudinal zonation variables: distance to source [km], mean discharge [m3/s], channel slope [m/m], water temperature [◦C]. Habitat structure variables: phytal-m, phytal-e, lithophytal, xylal, macropelal, micropelal, argillal, psammal, akal, mesolithal, macrolithal [% coverage of channel bottom]

Nutrients and eutrophication variables: N tot, o-PO4–P [mg/l], pH, oxygen saturation [abs(O2%–100%)]. − 2− + + 2+ 2+ Salinity variables: Cl ,SO4 ,Na ,K ,Mg ,Mg [mg/l], carbonate hardness [mmol/l]. 122 U. Braukmann, D. Böhme / Limnologica 41 (2011) 113–124

Despite a comparable organic pollution and similar nutrient Table 7 contents the zoocoenosis of the section at Breitungen and Barch- Reference conditions of the lower section of the River Werra regarding physico- chemical aspects of water quality (presumed potential natural conditions) (from feld is totally different from the one at Gerstungen. The Werra at Hübner 2007, modified). the sites at Breitungen and Barchfeld for instance may be charac- terised as a moderately polluted larger carbonate river of the central Parameter Reference condition low mountain range being inhabited by a species-rich fauna con- Chloride <100 mg/l sisting of characteristic benthic groups like mollusca, gammaridae, Sulphate Clearly below 100 mg/l Potassium ≤5 mg/l ephemeroptera, plecoptera, trichoptera, coleoptera und diptera in Calcium 60–80 mg/l relatively balanced proportions. Magnesium 10–15 mg/l The investigations of the invertebrate fauna of the Werra and Total hardness 2.0–2.5 mmol/l (11–14◦ Gh) its main tributearies carried out by Hübner (2002, 2005, 2007), Water temperature (summer) <19 to <20 ◦C Braukmann and Hübner (2003), Hübner and Braukmann (2005a,b, Saprobic index 1.65 Turbidity Low 2006) and in 2008 by Braukmann and collaborators (published here) reveal the fundamental faunistic change between Barchfeld and Gerstungen. Our comparative ecological investigations of the non-salt- At the site at Barchfeld 14 mostly river-specific, partly very contaminated Werra and the heavily polluted section as well as rare red-list-species, were found, e.g. Brachyptera braueri, numer- its tributaries (Hübner 2002, 2005, 2007; Braukmann and Hübner ous larval specimen of the caddisfly Brachycentrus subnubilus and 2003; Hübner and Braukmann 2005a,b, 2006) prove the fundamen- abundant specimen of the groundbug Aphelocheirus aestivalis.At tal change in the benthic community of macroinvertebrates. The Barchfeld altogether 74 different macroinvertebrate taxa were findings of the statistical analyses of the data collected in 2003 can found. be assumed to be valid until now. A comparison of multimetric In a sharp contrast to this finding the benthic invertebrate evaluation results for the sampling sites near Barchfeld, Gerstun- coenosis of the salt contaminated lower Werra at Gerstungen with gen and Blickershausen from 2003 (Table 3) and 2008 (Table 4) the highest salt concentration is extremely poor in species and reveals that the degradation of the benthic community within the nearly completely inhabited by alien species. It is mainly dom- salt polluted Werra section is as severe as some years ago, even inated by two neozoic salt-tolerant species Gammarus tigrinus though the metric “diversity” of the unpolluted site at Barchfeld which was introduced into the Werra by Schmitz in 1957 (Schmitz has slightly improved. 1960) and Potamopyrgus antipodarum, an invasive neozoon from New Zealand. All other species or taxa were found just in single Importance of the creek Salzbach with its natural salt composition specimens usually downstream from the mouth of freshwater tributaries. This observation is strongly related to the seasonality of the The Salzbach stream is a typical example of low magnesium benthos data, in detail to the deviant community pattern in spring. and potassium concentrations) in naturally salty running waters During snow melt the Werra as well as the tributaries enhance where parts of indigenous species like Gammarus pulex can survive flow rates and—at least temporarily-intensified transport of par- even under extreme sodium chloride conditions comparable to the ticulate matter. The salt load in the Werra is more diluted than in heavily salinised Werra section at Gerstungen. later seasons. Specimens of native species, drifting from unpolluted upstream river sections and tributaries into the salt polluted Werra Leitbild (reference conditions) for the River Werra section, have the chance to outlast there for days or weeks. How- ever, one may assume that the ostensibly successful colonisers are Based on his investigations, Hübner (2007) has reconstructed not able yet to complete here their reproductive cycle. They seem reference conditions (the “leitbild”) for the river Werra in its to establish no persistent populations and disappear throughout potentially non-salt-polluted lower sections, based on studies of the following months. historical data as well as recent research programmes carried In the area with the highest salt concentration nearly all typical out at the Department of Water Ecology and Management of the benthic groups named above disappeared. University of Kassel. These reference conditions represent the pre- At Gerstungen just 13 Taxa were observed (18% of the site at industrial status of the river’s chemistry and biology without salt Barchfeld). Only 9 out of 78 taxa in total match at both sampling pollution by the potash mining industry. Hence they provide a set sites (Barchfeld and Gerstungen). In terms of the quantitative Stein- of target values for the future restoration of the water body. These haus index, the similarity between these two sites is only 8.9%. reference conditions are summarised in Table 7. The sampling of macroinvertebrates carried out during pessimal According to this “leitbild” the River Werra should be naturally a quality conditions in August 2008 (low water level, high tempera- normal freshwater river of a carbonate type with an electrical con- tures, high fluctuation in oxygen concentration etc.) resulted in 34 ductivity of about 600 ␮S/cm, chloride concentration of clearly less taxa at Breitungen and 42 at Barchfeld. than 100 mg/l and natural potassium values less than 5 mg/l. These The total list of taxa (see Table 8, Appendix) gives an overview of conditions differ drastically from the actual chemical situation and the invertebrate assemblages of all sites of survey phase 2 in 2008. lead to severe damages of aquatic life in this salt polluted river. At Barchfeld there have been very little changes in the species stock As a result of the discharge of salt waste water the natural chem- between 2003 and 2008. The same species were dominant in both ical character of the Werra changes from a typical freshwater river years. Again a large individual number of Aphelocheirus aestivalis having a salinity of 0.4‰ to a miohaline (Remane 1971) river with and Brachycentrus subnubilus was found. Also both characteristic brackish water of 4.6‰ salinity, containing unnatural salt compo- river gammarids Gammarus pulex and Gammarus roeselii occurred sition and an extremely high chemically unbalanced total hardness in high numbers. of 70◦ Gh (12.5 mmol/l). The salt concentration increases, and this A considerable drop in taxa number from 42 to just 4 took place is caused by direct discharge and indirect diffuse salt ingress in at Gerstungen, the site with the highest salt concentration. The the area of the grouting (Verpressung) between Hattorf and Ger- saltwater gammarid Gammarus tigrinus was absolutely dominant stungen by 11.5-fold. Potassium and magnesium concentrations (98.4% dominance). Other taxa were: Potamopyrgus antipodarum are extremely overrated due to the insufficient separation from the (1.4%), Dendrocoelum lacteum (0.1%) and Oulimnius rivularis (1 spec- waste salt components compared to the natural situation in the imen, 0.03%). Werra. U. Braukmann, D. Böhme / Limnologica 41 (2011) 113–124 123

The continuing high salt load prevents a permanent recolonisa- potassium and magnesium concentrations. Thus the actual legal tion by typical sensitive freshwater species. Above all the toxical limits for chloride and total hardness have to be revised. impact of potassium on freshwater invertebrates like gammarids has been proved sufficiently in literature for a long time (Albrecht 1954; Schmitz et al. 1967; Koop 1994). Regarding water man- Acknowledgments agement aspects, it is necessary to reduce not only the chloride concentration of the Werra from 2500 mg/l to less than 1000 mg/l Many thanks to Petra Knoch-Hüser who managed the chemi- but especially the concentrations of potassium from currently more cal analyses in the lab and to Bastian Rupp and Ulf Stein for the than 150 mg/l by a factor of 10 and magnesium from about 220 mg/l assistance in the field. Thanks also to Dr. Gerd Hübner whose by a factor of 15 to at least 15 mg/l each. Further utilisation of the unpublished precise original data on macroinvertebrates (2003) actual legal limits by the potash mining industry obviously impedes collected in our team could be used as a valuable base for this paper. a considerable improvement of the ecological status of the Werra. This obligatory objective is demanded by the European WFD. It can Appendix A. be achieved only by a decrease of the overall salt load in combi- nation with a significant reduction of the physiologically harmful Table 8

Table 8 Taxa list of four sampling sites of the River Werra, August 2008, Ind/m2.

Taxaname Breitungen Barchfeld Gerstungen Blickershausen

Baetis fuscatus 700 7 Simulium sp. 200 10 Baetis vardarensis 200 3 Serratella ignita 55 58 Elmis maugetii 35 45 Leuctra fusca-Gr. 30 25 Hydropsyche incognita 21 6 49 Aphelocheirus aestivalis 10 210 Dugesia gonocephala 10 15 7 Hydroptila sp. 10 5 15 Lasiocephala basalis 10 5 Sphaerium corneum 920 Rhyacophila nubila-Gr. 9 9 Orthocladiinae 8 Atherix ibis 65 5 Brachycentrus subnubilus 5 125 Stylodrilus heringianus 52 Gammarus roeselii 460 Simulium argenteostriatum 44 Ancylus ﬂuviatilis 41 Gammarus pulex 3 120 Heptagenia sulphurea 310 Piscicola geometra 32 Perlodes sp. 3 Wilhelmia-Gr. 2 2 Lumbriculidae 2 1 10 Radix balthica 12 3 Erpobdella octoculata 12 Brillia cf. longifurca 11 Eiseniella tetraedra 11 Helobdella stagnalis 11 Tanypodinae 1 1 Brychius elevatus 1 Goeridae 1 Barbatula barbatula 1 Asellus aquaticus 20 Bithynia tentaculata 10 Potamophylax rotundipennis 7 Dendrocoelum lacteum 35 3 Prodiamesa olivacea 124 Ceraclea albimacula-alboguttata-Gr. 1 Chironomus sp. 1 Colymbetes fuscus 1 Ecdyonurus torrentis 1 Ephemera danica 1 Halesus tesselatus 1 Plectrocnemia sp. 1 Gasteroceus aculeatus 1 Gammarus tigrinus 3500 2000 Potamopyrgus antipodarum 50 1000 Dicrotendipes sp. 95 Proasellus coxalis 65 Hydropsyche contubernalis 14 124 U. Braukmann, D. Böhme / Limnologica 41 (2011) 113–124

Table 8(Continued)

Taxaname Breitungen Barchfeld Gerstungen Blickershausen

Tubiﬁcidae 5 Limnius volckmari 2 Anacaena bipustulata 1 Chironomus plumosus-Gr. 1 Dicrotendipes notatus 1 Nepa cinerea 1 Oulimnius cf. rivularis 1

References Legendre, P., Legendre, L., 1998. Numerical Ecology. Second English Edition. Elsevier Science B.V., Amsterdam. Albrecht, M.-L., 1954. Die Wirkung der Kaliabwässer auf die Fauna der Werra und Lorenz, A., Hering, D., Feld, C.K., Rolauffs, P., 2004. A new method for assessing the Wipper. Zeitschrift für Fischerei und deren Hilfswissenschaften 3, 401–426. impact of hydromorphological degradation on the macroinvertebrate fauna in Asterics/Perlodes, 2008. European stream assessment program, system perlodes. five German stream types. Hydrobiologia 516, 107–127. Contact No. EVK1-CT1999-00027, Version 3.1.1. Mantel, N., 1967. The detection of disease clustering and a generalized regression Braukmann, U., 2000. Hydrochemische und biologische Merkmale regionaler approach. Cancer Res. 27, 209–220. Bachtypen in Baden-Württemberg. Landesanstalt für Umweltschutz Baden- Moog, O., 1995. Fauna aquatica austriaca. Katalog zur autökologischen Einstu- Württemberg. Oberirdische Gewässer, Gewässerökologie 56, Karlsruhe. fung aquatischer Organismen Österreichs. Bundesministerium für Land-und Braukmann, U., Hübner, G., 2003. Gewässerökologische Forschung an der Werra und Forstwirtschaft. Loseblattsammlung, Wien. die Wasserrahmenrichtlinie der Europäischen Union. Jahrbuch Naturschutz in Pottgiesser, T., Sommerhäuser, M., 2004. Fließgewässertypologie Deutschlands. Hessen 8, 17–30. Die Gewässertypen und ihre Steckbriefe als Beitrag zur Umsetzung der Clarke, K.R., 1993. Non-parametric multivariate analysis of changes in community EU-Wasserrahmenrichtlinie. In: Steinberg, C., Calmano, W., Klapper, H., structure. Aust. J. Ecol. 18, 117–143. Wilken, R.-D. (Eds.), Handbuch Angewandte Limnologie, Verl. ecomed. 19. Hering, D., Moog, O., Sandin, L., Verdonschot, P.F.M., 2004. Overview and application Ergänzungslieferung 7/2004, VIII–2.1, 3–16+Anh. A+B, Landsberg/Lech. of the AQEM assessment system. Hydrobiologia 516, 1–20. Pottgiesser, T., Sommerhäuser, M., 2008. Begleittext zur Aktualisierung der Steck- Hübner, G., 2002. Die historische Fischfauna der unteren Werra. Philippia 10 (2), briefe der bundesdeutschen Fließgewässertypen (Teil A) und Ergänzung 119–129. der Steckbriefe der deutschen Fließgewässertypen um typspezifische Ref- Hübner, G., 2005. Vergleichende Untersuchung zur Köcherfliegen-Besiedlung der erenzbedingungen und Bewertungsverfahren aller Qualitätselemente (Teil salzbelasteten unteren Werra (Weser) und ihrer Hauptzuflüsse. Lauterbornia B). UBA-Projekt (Förderkennzeichen 36015007) und LAWA-Projekt O 8.06. 54, 91–102. www.wasserblick.net. Hübner, G., 2007. Ökologisch-faunistische Fließgewässerbewertung am Beispiel Remane, A., 1971. Ecology of brackish water. In: Remane, A., Schlieper, C. (Eds.), Biol- der salzbelasteten unteren Werra und ausgewählter Zuflüsse. Doctoral thesis ogy of Brackish Water, second ed. Die Binnengewässer 25. Verl. E. Schweizerbart, Universität Kassel, Fachbereich Architektur, Stadtplanung, Landschaftsplanung. Stuttgart, pp. 1–210. Ökologie und Umweltsicherung, 27/2007, Witzenhausen. Richardson, R.E., 1928. The bottom fauna of the Middle Illinois River, 1913–1925. Ill. Hübner, G., Braukmann, U., 2005a. Referenzorientierte ökologische Bewertung der Nat. Hist. Surv. Bull. 17, 387–475. salzbelasteten Werra anhand des Makrozoobenthos. In: Deutsche Gesellschaft Schmedtje, U., Colling, M., 1996. Ökologische Typisierung der aquatischen Makro- für Limnologie (DGL). Tagungsbericht Jahrestagung Potsdam 2004, pp. 120–124. fauna. Informationsberichte des Bayerischen Landesamtes für Wasserwirtschaft Hübner, G., Braukmann, U., 2005b. Zur Entwicklung der Süßwasser-Mollusken (Gas- 4/96, München. tropoda et Bivalvia) im Flussgebiet der unteren Werra. Philippia 12 (2), 125–136. Schmitz, W., 1960. Die Einbürgerung von Gammarus tigrinus Sexton auf dem Hübner, G., Braukmann, U., 2006. Leitbild, Referenz und europäischen Kontinent. Arch. Hydrobiol. 57 (1/2), 223–225. Wiederbesiedlungspotenzial–ein Bewertungsrahmen für Erfolgskontrollen Schmitz, W., Besch, W., Kneisel, I., 1967. Die Salzgehaltstoleranz von Gammarus pulex an der salzbelasteten unteren Werra. Informationsdienst Naturschutz pulex (L.), Gammarus tigrinus Sexton und Asellus aquaticus (L.) in Abhängigkeit Niedersachsen 26 (2), 137–144. von der relativen Konzentration der Kationen Na, Mg, K und Ca. Int. Rev. Ges. Koop, J., 1994. Akute und subakute Einflüsse von Kaliumchlorid-Belastungen in Hydrobiol. 52 (4), 589–616. Fließgewässern auf den Energiestoffwechsel von Gammarus pulex (L.) und Gam- Sneath, P.H.A., Sokal, R.R., 1973. Numerical Taxonomy—The Principles and Practice marus tigrinus (Sexton). Doctoral Thesis Universität Düsseldorf, Berichte aus der of Numerical Classification. W.H. Freeman, San Francisco. Biologie, Verl. Shaker. WFD (Water Framework Directive), 2000. European Commission. Directive 2000/EC LAWA, 2007. RAKON, Arbeitspapier II (Rahmenkonzeption Monitoring, Teil of the European Parliament and the Council establishing a framework for com- B) Hintergrund-und Orientierungswerte für physikalisch-chemische munity action in the field of water policy. PE-CONS 3639/00, Bruxelles. Komponenten—Bewertungsgrundlagen und Methodenbeschreibungen), Stand 7.03.2007.